Odette> Wondering how can I generate "Akaike weight" with R? I know the description, > but is there any function to generate by R on the web-site or R library? > I am using GLM or GLMM (family=binomial), so would be appreciated if you > help me.You could have a look at this. http://bm2.genes.nig.ac.jp/RGM2/R_current/library/aod/man/summary.aic.html Which is in the OAD package Graham
I'm pretty sure you have to work it out yourself. Here is an example of how
you would tabulate the AIC weights from three models (check that my
calculations are correct before using this yourself!).
Basically model.name$aic will cut out the AIC values then write a formula to
calculate the weights...
m1<-glm(trantot~unimpgrass+impgrass,family=poisson,data=bbs[bbs$species==unique(bbs$species)[1],])
m2<-glm(trantot~marginwidth,family=poisson,data=bbs[bbs$species==unique(bbs$species)[1],])
m3<-glm(trantot~impgrass,family=poisson,data=bbs[bbs$species==unique(bbs$species)[1],])
aics<-data.frame(paste("m",1:3,sep=""),c(m1$aic,m2$aic,m3$aic),row.names=NULL)
colnames(aics)<-c("model","AIC")
aics<-aics[order(-aics$AIC),]
for(i in 1:dim(aics)[1]){
aics$diff[i]<-aics$AIC[1]-aics$AIC[i]}
aics$wi<-2.71828182845904523536^(-0.5*aics$diff)
aics$aic.weights<-aics$wi/sum(aics$wi)
----- Original Message -----
From: "Odette Gaston" <odette.gaston at gmail.com>
To: <r-help at r-project.org>
Sent: Friday, December 19, 2008 11:26 AM
Subject: [R] Akaike weight in R
> Hi folks,
>
> Wondering how can I generate "Akaike weight" with R? I know the
> description,
> but is there any function to generate by R on the web-site or R library?
> I am using GLM or GLMM (family=binomial), so would be appreciated if you
> help me.
> Thanks for your contributions in advance,
>
> Regards,
> Odette
>
> [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
Try AICtab in the bbmle package> ----- Original Message ----- > From: "Odette Gaston" <odette.gaston at gmail.com> > To: <r-help at r-project.org> > Sent: Friday, December 19, 2008 11:26 AM > Subject: [R] Akaike weight in R > > >> Hi folks, >> >> Wondering how can I generate "Akaike weight" with R? I know the >> description, >> but is there any function to generate by R on the web-site or R library? >> I am using GLM or GLMM (family=binomial), so would be appreciated if you >> help me. >> Thanks for your contributions in advance, >> > >-- View this message in context: http://www.nabble.com/Akaike-weight-in-R-tp21089767p21091392.html Sent from the R help mailing list archive at Nabble.com.
Seems pretty clear from a quick glance at that page (and in retrospect, even from the URL itself) that Graham misspelled the package name, Try looking for package, aod (without any caps.) -- David Winsemius On Dec 21, 2008, at 5:17 AM, Odette Gaston wrote:> Hi Graham, > > Appreciate for your help, but actually I've already looked at this > website, > and wasn't success yet. 1st, I cannot run "*W = exp(-0.5 * Delta) / > sum(exp(-0.5 * Delta))*" in the R. > >> *W <- exp(-0.5 * Delta) / sum(exp(-0.5 * Delta))* > *note: Delta means AIC difference between models.* > ** > I don't know why but always showing 1 even I changed Delta. > > Also, I couldn't find OAD package, where can I get one? > Would be wonderfully appreciated. > > Thanks in advance, > Odette > > > > > On Fri, Dec 19, 2008 at 8:33 PM, Graham Smith <myotisone at gmail.com> > wrote: > >> Odette >> >>> Wondering how can I generate "Akaike weight" with R? I know the >> description, >>> but is there any function to generate by R on the web-site or R >>> library? >>> I am using GLM or GLMM (family=binomial), so would be appreciated >>> if you >>> help me. >> >> You could have a look at this. >> >> http://bm2.genes.nig.ac.jp/RGM2/R_current/library/aod/man/summary.aic.html >> >> Which is in the OAD package >> >> Graham >> > > [[alternative HTML version deleted]] > > ______________________________________________ > R-help at r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code.
> Seems pretty clear from a quick glance at that page (and in retrospect, even > from the URL itself) that Graham misspelled the package name, Try looking > for package, aod (without any caps.)Yep, mea culpa (and I have probably spelt that wrong as well), it was quick after thought as I almost just gave the link to the page. Sorry for the confusion. Graham